http api like the one i made

badger has a good read cache on it as well, and it's designed to split the read/write load mainly in favor of writing indexes because events don't change, also, scanning the key tables is fast because it's not interspersed with value read/write